Improvement in accuracy of evaluating the customer service of operators in a call center, etc., has been sought in recent years. Conventionally, evaluation of situations of customer service of operators has been made by listening to the call between the customer and the operator later and by evaluating the operator's response to the customer. For that reason, at the time of evaluating an administrator and an operator are not able to quickly determine if there is a problem in the customer service of operators.
For example, a technology to evaluate a call of an operator in a call center by using utterances of the operator and a customer has been known. In this technology, technologies in which audio input from an operator or a customer undergoes audio processing (power or utterance section and silent section) and the processed audio is output, and a warning is issued as a result of a comparison with a certain threshold, have been known. For example, when it is determined that the operator's utterances include too many particular expressions such as interjections or too little explanatory vocabulary, or is too loud, too quiet, or too fast, or when it is determined that a ratio of one speaker's utterances to another speaker's silence is large, a warning is issued.
Moreover, a technology to compare a ratio of utterances between an operator and a customer and to determine whether or not the ratio of utterances is biased by using audio information of an operator and a customer in a prescribed period of time during a call has been known. When the ratio is determined to be biased, an administrator who is in charge of the operator or the operator receives the warning against improper customer service of the operator. Furthermore, a technology that obtains an utterance ratio between an operator and a customer and displays the utterance ratio on a terminal of an administrator in a chronological order has been known.
In addition, a technology to evaluate the content of answers of an operator so as to provide quick and accurate answers to a customer's questions has been known. According to this technology, an evaluation value is obtained on the basis of a predetermined evaluation formula by using the speed of an operator from receiving a question to an output of an answer e-mail to a customer or the length of the answer provided in the answer e-mail, and an evaluation is made on the basis of the evaluation value.
It has also been known that by storing questions from customers in the past and an operator's answers to the questions in association with customer's evaluations of the answers, when a similar question is received, an operator who is highly evaluated for the answer to the question is assigned.
Furthermore, a technology to graphically display an order of uttering in a remote conference system has been known. According to this technology, utterance sections for each speaker are displayed in a chronological order by analyzing overlapping or continuity of utterance sections by using audio information for each speaker.
Patent Document 1: Japanese Laid-Open Patent Publication No. 2007-33754
Patent Document 2: Japanese Laid-Open Patent Publication No. 2009-216840
Patent Document 3: Japanese Laid-Open Patent Publication No. 2001-297259
Patent Document 4: Japanese Laid-Open Patent Publication No. 2004-157733
Patent Document 5: Japanese Laid-Open Patent Publication No. 2007-58767